Content uploaded by Gilla Shapiro
Author content
All content in this area was uploaded by Gilla Shapiro on Oct 16, 2014
Content may be subject to copyright.
Measuring Financial Anxiety
Gilla K. Shapiro
London School of Economics and Political Science Brendan J. Burchell
University of Cambridge
There is a scarcity of information concerning the emotional aspects of financial
management. Two studies were conducted to evaluate the measurement of conscious
and intuitive emotional anxiety toward one’s personal finances. Along with a self-
reported financial anxiety questionnaire, a modified Emotional Stroop Test (EST) and
Dot-Probe Paradigm (DPP) were separately utilized to evaluate financial anxiety. In
both studies, the self-reported financial anxiety questionnaire correlated significantly
with the implicit measures. Furthermore, the DPP was predominantly characterized by
avoidance of financial information. Financial anxiety was shown to be a separate
construct from depression and general anxiety. These findings indicate that those who
report having financial anxiety also display reaction latencies in the processing of
financial information. Accordingly, financial behavior could be more comprehensively
evaluated and policy could be better determined by incorporating financial anxiety into
models of financial illiteracy, mismanagement, and debt.
Keywords: financial anxiety, dot-probe paradigm, emotional Stroop test, student debt, financial
policy
Supplemental materials: http://dx.doi.org/10.1037/a0027647.supp
The ability to make informed and effective
decisions regarding the management of money
is important for individual success (Jorgensen,
2007). The requirement for financial compe-
tence is particularly essential in the current fi-
nancial climate. Instead, individuals display
surprisingly poor financial literacy and a wide-
spread lack of knowledge on fundamental eco-
nomic concepts (Jump$tart Coalition, 2004;
NCEE, 2005; Warwick & Mansfield, 2000),
patterns of overspending (Roberts & Jones,
2001), and careless financial behavior (Henry,
Weber, & Yarbrough, 2001). These problems
are interlinked and financial illiteracy has been
associated with both financial mismanagement
(Hastings & Tejeda-Ashton, 2008) and debt
(Lusardi & Tufano, 2009). The combined result
is that individuals find themselves in substantial
debt that can leave them seriously indebted for
years, situated in a “debt cycle” (Beal & Del-
phachitra, 2003; Wang & Xiao, 2009).
Debt is not only problematic in itself, but is
related to negative psychological repercussions
including a decreased sense of ability to manage
one’s money, lower self-esteem, decreased
sense of financial wellbeing, lower productivity,
and higher levels of overall stress (Garman,
Leech, & Grable, 1996; Joo & Grable, 2000;
Lange & Byrd, 1998). This is exemplified by
Roberts, Golding, Towell, and Weinreb’s
(1999) finding that British students who have
considered leaving university for financial rea-
sons were more likely to report poorer mental
health and social functioning. Furthermore, in a
nationally representative household sample,
Jenkins et al. (2008) found that 23.8% of people
with a mental disorder (i.e., neurotic disorder,
psychotic disorder, alcohol dependence, or drug
dependence) were in debt compared to 8.1% of
people without these disorders, and the more
debts people incurred the more likely they were
to have a mental disorder. Future research
This article was published Online First March 12, 2012.
Gilla K. Shapiro, MPA Programme, Institute of Public
Affairs, London School of Economics and Political Science,
London, England; Brendan J. Burchell, Department of So-
ciology, University of Cambridge, Cambridge, England.
This work was conducted while the first author was a
student at the Department of Psychology, University of
Cambridge. We are grateful for the critical comments and
input given by Dr. Luke Clark.
Correspondence concerning this article should be ad-
dressed to Gilla K. Shapiro, MPA Programme, Institute of
Public Affairs, London School of Economics and Political
Science, Houghton Street, London, England, WC2A 2AE.
E-mail: gilla.shapiro@cantab.net
Journal of Neuroscience, Psychology, and Economics © 2012 American Psychological Association
2012, Vol. 5, No. 2, 92–103 1937-321X/12/$12.00 DOI: 10.1037/a0027647
92
would be beneficial in determining the causal
pathways of debt and mental illness.
The combination of financial illiteracy and
mismanagement denotes that a substantial num-
ber of individuals face the foreseeable risk of
considerable debt and associated psychological
difficulties (Warwick & Manfield, 2000). Re-
searchers and policymakers have generally as-
sumed that a lack of knowledge is responsible
for financial illiteracy and mismanagement.
Contrary to this view, there is beginning to be
awareness of the important emotional compo-
nent to effective decision making and financial
competence (Burchell, 2003; Hanoch, 2002).
While initially it became recognized through
neurobiological studies that an absence of emo-
tions leads to suboptimal decisions (Damasio,
1994), psychological research has confirmed
that strong emotional responses are also associ-
ated with poor decisions (particularly those of a
financial nature) (Ackert, Church, & Deaves,
2003). As Rolls (1999) explains, positive feel-
ings improve individuals’ ability to organize
and assimilate information, problem-solve, ne-
gotiate, and make efficient decisions. Burchell’s
(2002) combination of in-depth interviews fol-
lowed by a telephone survey of 1,000 adults
suggested that avoidant reactions were specific
to personal finances, and did not necessarily
spill over into other domains such as financial
responsibilities in the workplace. Furthermore,
Burchell (2002) found that out of three psycho-
metrically sound subscales measuring “knowl-
edge,” “importance” and “emotion” toward fi-
nances, the emotional scale was the best predic-
tor of financial behavior.
Based on this emerging research, it seems
plausible that a strong negative emotional re-
sponse of financial anxiety is associated with
financial illiteracy and mismanagement. Finan-
cial anxiety has been defined as a psychosocial
syndrome whereby individuals have an uneasy
and unhealthy attitude toward engaging with,
and administering their personal finances in an
effective way (Burchell, 2003); this study gave
rise to the term financial phobia being widely
discussed in popular media. However, financial
anxiety, despite having important repercussions
for understanding financial behavior, remains
largely understated and uninvestigated. The
purpose of the two studies described below was
to add to the relatively meager literature by
investigating financial anxiety through develop-
ing a self-report questionnaire (FAS) and adapt-
ing intuitive tools (the Emotional Stroop Test
and the Dot-Probe Paradigm) to measure this
phenomenon.
Methodological Note
The measurement of individuals’ financial
anxiety toward dealing with their personal fi-
nances remains largely uninvestigated and no
tools (to the authors’ knowledge) currently
measure this phenomenon. However, tools de-
veloped in recent years and applied to the mea-
surement of other phobias might be usefully
employed to measure financial anxiety.
The construct of financial anxiety developed
from investigations of financial attitudes (Lim
& Sng, 2006). It is therefore not surprising that
the few self-report measures that test for finan-
cial anxiety are subscales developed to measure
more general, “multidimentional” money atti-
tudes. The two foremost money-attitudes mea-
sures are Yamauchi and Templer’s (1982)
Money Attitude Scale (MAS), and Furnham’s
(1984) Money Beliefs and Behavior Scale
(MBBS). There is encouraging overlap between
these scales (Roberts & Jones, 2001), but while
the longer (60-item) MBBS has been criticized
for low reliability, the (29-item) MAS has in-
stead shown stable factor structure and loadings
in multinational and ethnically diverse samples
(Baker & Hagedorn, 2008; Medina, Saegert, &
Gresham, 1996). However, neither of these anx-
iety subscales would be applicable to studying
financial anxiety because questions on these
subscales were consumer specific (e.g., “I am
bothered when I have to pass up a sale”) and
inappropriate for measuring financial anxiety
among all ages (e.g., “I save now to prepare for
my old age”). Only Burchell (2003) has created
a self-report (5-item) scale measuring general
emotional orientation of individuals toward
their personal finances.
That emotional responses have been associ-
ated with psychophysiological signals (Dama-
sio, 1994) indicates that financial anxiety can be
measured by subconscious measures as well as
conscious, subjective reports. Researchers have
studied emotion’s interference with attention
through utilizing different variants of the Stroop
task or the Dot-Probe Paradigm. In order to
measure financial anxiety, the two studies pre-
sented in this paper developed a self-reported
93MEASURING FINANCIAL ANXIETY
Financial Anxiety Scale (FAS) and adapted two
well-established implicit instruments (i.e., Emo-
tional Stroop Test and Dot-Probe Paradigm),
which have not hitherto been applied to mea-
suring financial anxiety. These tools were de-
veloped to measure financial anxiety and inves-
tigate how reported financial anxiety is related
to the involuntary processing bias of financial
anxiety.
As Miu, Miclea, and Houser (2008) explain,
the theoretical and methodological changes in
studying emotions have stemmed from multi-
disciplinary efforts bridging psychology, phys-
iology and neuroscience. These studies extend
this works in its methodology of measuring
financial anxiety.
Study 1
The first study utilized the Emotional Stroop
Test (EST), which is commonly used to dem-
onstrate the existence of attentional biases in a
wide range of psychological disorders (Wil-
liams, Macleod, & Mathews, 1996). In the EST,
emotionally relevant (“threat”) words have been
found to capture the attention of the individual
and cause an attentional bias toward threatening
information (Amir et al., 1996; Williams et al.,
1996).
Though a popular methodology to test a wide
range of anxieties, the EST has not been
adapted to measure financial anxiety. This study
adapted the EST to measure a subliminal bias to
financial words and developed a financial anx-
iety scale (FAS) to measure self-reported finan-
cial anxiety. This study seeks to evaluate
whether a reported financial anxiety is related to
a subliminal bias in processing financially
loaded words.
Hypothesis. It was hypothesized that there
would be a negative relationship between
the difference in RTs of the financial and
neutral words on the FEST with the self-
reported responses on the Financial Anxi-
ety Scale (FAS).
1
Method
Participants. The participants in this study
were 38 unpaid volunteers, 22 males and 16
females, 19–22 years of age [M ⫽20.17, SD ⫽
.78], who were full-time undergraduate students
responding to fliers and emails. Similar to Ro-
fey, Corcoran, and Tran (2004), this study ex-
cluded those participants who were colorblind
as this would have been problematic on the
Emotional Stroop Test, as well as those who are
not fluent in English.
Instruments and measures.
Financial Emotional Stroop Test (FEST).
The Financial Emotional Stroop Test (FEST)
used eight words that were presented twice in
four different colors (as in the original Stroop):
red, green, blue, and yellow (see Table 1). The
four different colors must be read aloud, while
ignoring the word that is colored. For example,
the word overdraft colored in blue must be read
as blue, and the presentation of the word over-
draft should be ignored.
There are three groups of words in the FEST:
(1) a “classic condition” whereby classic incon-
gruent color words were presented, (2) a “finan-
cial condition” with financially loaded words,
and (3) a “control condition” with neutral words
that correspond to the financial condition (see
Table 1). The classic words were based upon
the original Stroop Test (Stroop, 1935), whereas
the financial group of words and their neutral
counterparts were designed for this experiment.
Using a computer ST opposed to a card ST
was disregarded based on Kindt, Bierman, and
Brosschot’s (1996) study that showed “the
highest test–retest correlation for the standard
Stroop effect was found on the card format”
(p.659). As specified in the German adaptation
of the Stroop, participants in this study were
instructed to read the colors of the words col-
umn-wise as quickly and accurately as possible
(Baumler et al., 1985 as reported by Stetter,
Ackermann, Bizer, Straube, & Mann, 1995).
Word length and syllables in the financial and
neutral-control group of words were matched so
that word length and subvocalized reading was
controlled (see Table 1).
Financial Anxiety Scale (FAS). The
(FAS) measures an anxious disposition toward
cognitive engagement with one’s personal fi-
1
A greater difference in RTs of the financial and neutral
words on the FEST indicates greater financial anxiety, while
a greater score of the FAS indicates less financial anxiety.
94 SHAPIRO AND BURCHELL
nances (see Table 2).
2
Some of the questions
were borrowed from Burchell (2003) and Da-
vies and Lea (1995), while the authors formu-
lated other questions. Lower numbers indicate a
higher level of financial anxiety on this 4-point
Likert scale, which ranged from 1 very true to 4
completely untrue.
Procedure. The FAS was administered
following the FEST. The neutral and financial
words in the FEST were counterbalanced and
no learning effect was found.
3
3.2. Results
The data was analyzed using SPSS ver-
sion 16.0 for MacOS X (SPSS Inc., 1989–
2007). On average, the student population re-
ported having a mean score of 2.9 on the FAS
(SD ⫽0.64).
4
In comparing the response times on the
FEST, the highest latency was recorded for the
classic condition [M⫽11.24 s, SD ⫽1.83]. On
average the financial condition had a slightly
slower response time [M⫽8.48 s, SD ⫽1.45]
compared to the control condition [M⫽8.30 s,
SD ⫽1.25]. A paired sample ttest was con-
ducted to see if the mean of the financial con-
dition is significantly greater than the mean of
the neutral condition. The two-tailed ttest re-
vealed that these conditions means are not sig-
nificantly different, t(37) ⫽⫺1.115, ns.
Thedifferentialoftheneutralresponsetimeand
thefinancial response time was calculated for each
participant, which was then correlated with each
participant’s FAS score. A one-tailed Pearson’s
product-moment correlation between the interface
response time and FAS confirmed a negative cor-
relation (r ⫽⫺.308), which was significant (p⫽
.05) and had a medium effect size.
Discussion
The results indicate that those who reported
having higher levels of financial anxiety (on the
FAS) also displayed a reaction latency on finan-
cially loaded words. This significant correlation
indicates that a reported financial anxiety is
related to a subliminal bias in processing finan-
cially loaded words.
While De Ruiter and Brosschot (1994) have
argued that vigilance and avoidance processes
may be detectable on the EST, other researchers
have suggested that it is difficult to interpret
whether the EST displays vigilance or avoid-
ance to threat words because the unattended
stimuli are spatially separate (as the financial
and neutral words were tested and scored sep-
arately) (Fox, 1993; MacLeod, Mathews, &
Tata, 1986).
2
An exploratory factor analysis found that three of the
original variables included in the questionnaire were not
associated with the same factor. This was confirmed by
inspecting the correlation matrixes. These three questions
were therefore excluded from the FAS (Table 2). The ex-
clusion of these items increased the Cronbach’s alpha (␣⫽
.809 to ␣⫽.828).
3
Independent samples t-tests were run to see if the means
of the neutral and financial conditions were significantly
different between the two order-effect groups. For both the
financial and neutral condition, it was found that the differ-
ence was not significant [t
Financial
(36) ⫽.085, ns; t
Neutral
(36) ⫽⫺.413, ns].
4
An average score of 1 indicates greatest possible anxiety,
while a score of 4 indicates least possible anxiety on the FAS.
Table 1
Groups of Words Used in FEST
Classic condition Financial condition Control condition
Orange Overdraft Navigates
Green Credit Apron
Brown Bank Shoe
Purple Loan Chair
Turquoise Cost Frog
Yellow Account Mantle
Red Debt Port
Blue Finances Cinemas
Total number of syllables 12 14 14
Total number of letters 44 46 44
Note. A color version of this table is available online as Supplementary Material.
95MEASURING FINANCIAL ANXIETY
Study 2
This study dealt with two potential criticisms
of Study 1. First, both low mood and a general
anxiety were not controlled in Study 1. As
symptoms of depression and anxiety are preva-
lent among the student population (Ricciardi,
2008), it is possible that the devised question-
naire was vicariously being contaminated by
these variables rather than financial anxiety.
This study aimed to verify if a financial anxiety
measure is distinct from more general symp-
toms of anxiety or depression.
Second, the previous study is a modified ver-
sion of the EST. Despite wide use, the EST has
been criticized as inconsistent, unable to di-
rectly compare threat and neutral words (Fox,
1993), and difficult to interpret (Morgan, Rees,
& Curran, 2008). In response to the problems
with the EST, MacLeod et al. (1986) developed
the Dot-Probe Paradigm (DPP), which im-
proves upon the EST as “it eliminates the pos-
sibility of response bias interpretations....
[and] it allows a test of the prediction that the
presence of a threatening term can both facili-
tate and impair dot detection, in the same indi-
vidual, depending on the threat word’s position
relative to the dot” (MacLeod et al., 1986, p.18;
Tata, Leibowitz, Prunty, Cameron, & Pickering,
1996). Accordingly, the design of the DPP en-
ables researchers to determine whether the
mechanism underlying financial anxiety is one
of vigilance or avoidance. The DPP has been
used on clinical and nonclinical samples
(Ehrman et al., 2002; Pflugshaupt et al., 2005;
Townshend & Duka, 2001). However, only
Morgan et al. (2008) have conducted the DPP
with money-related words, but the focus of
Morgan et al.’s study was the attentional bias of
ketamine users (and money words were used as
a nondrug incentive control). The DPP has not
been used to test financial anxiety. Study 2 was
therefore designed to retest and extend Study 1.
Table 2
Financial Anxiety Scale
#
Factor loading on component
1 before itemsa&dwere
deselected Factor loading on
component 1 Question
a° .438 — I find monitoring my bank or credit card accounts very
boring
b .736 .706 I prefer not to think about the state of my personal
finances
c .737 .740 Thinking about my personal finances can make me feel
guilty
d°• .373 —
There’s little point in saving money and being careful
with it, because you could lose it all through no fault
of your own
e• .486 .526 I am worried about the debt I will have when I complete
my university education
f .643 .687 Thinking about my personal finances can make me feel
anxious
g .642 .625 I get myself into situations where I do not know where
I’m going to get the money to “bail” myself out
h .693 .709 Discussing my finances can make my heart race or make
me feel stressed
i• .658 .652 I do not make a big enough effort to understand my
finances
j .572 .567 I do not think I am doing as well as I could
academically because I worry about money
k
ⴱ
.784 .782 I find opening my bank statements unpleasant
l
ⴱ
.603 .623 I would rather someone else who I trusted kept my
finances organized
Note.
ⴱ
New items incorporated into Study 2. • items that were excluded from Study 1’s analysis based on an exploratory
factor analysis and their Chronbach’s alpha; however, these were retested because Study 1 had a small sample size (N⫽
38). ° items that were excluded from Study 2’s analysis based on an exploratory factor analysis and their Chronbach’s ␣.
96 SHAPIRO AND BURCHELL
Hypothesis 1. It was hypothesized that fi-
nancial anxiety is a measurable phenome-
non that is distinct from depression and a
generalized anxiety.
Hypothesis 2. Following the results from
Study 1, it was hypothesized that a mod-
erate relationship between self-reported fi-
nancial anxiety (on the FAS) and implicit
measured anxiety (on the DPP) would be
replicated.
Hypothesis 3. It was hypothesized that
avoidance would be the mechanism in-
volved in financial anxiety (measured on
the DPP) rather than attentional capture.
Method
Participant recruitment and retention.
An a priori power calculation through G
ⴱ
Power
was conducted to investigate how many partic-
ipants were necessary for this study (Faul &
Erdfelder, 1992). It was possible to conduct this
analysis because parameter estimates could be
based on the results from Study 1. Using the
effect size (r⫽.308) and Alpha (0.05) from
Study 1, and recognizing that a power of 0.8 is
recommended (Field, 2005), 61 participants
were found to be necessary.
To allow for dropouts and erroneous data, 79
unpaid volunteers were recruited by responding
to fliers and emails. These were 40 males and 39
females between the ages of 18–22 [M⫽20.05,
SD ⫽1.19]. Recruitment requirements were
fluency in English and that participants were
full-time undergraduate students.
One participant was initially excluded be-
cause of a problem with their inputted data. A
further four participants were excluded because
they either had too many errors (less than 119
correct out of 136 trials), a mean reaction time
(RT) of less than 200 ms (indicating that the
subject consistently initiated a response before
the onset of the target), or a mean RT of greater
than 1,000 ms (indicating anomalous processing
such as inattention to task) (Townshend &
Duka, 2001). In total, 6.3% of participants were
excluded (n⫽5). There were no significant
differences in the excluded participants’ demo-
graphic characteristics.
Instruments and measures.
Questionnaires. Three self-reported ques-
tionnaires were administered. The FAS (used in
Study 1) measured 10 questions on a 4-point
Likert Scale, which were equally weighted in a
cumulative score (see Table 2) and then aver-
aged so to represent a score out of four. The
Spielberger’s State–Trait Anxiety Inventory
(STAI) and Center for Epidemiologic Studies
Depression Scale (CES-D), both scored on a
4-point Likert Scale, were respectively used to
evaluate whether higher anxiety or low mood
can account for financial anxiety. The STAI
measures feelings such as tension, nervousness,
and confusion (VanderZee, Sanderman, &
Heyink, 1996), while the CES-D addresses
items such as enjoyment of life, hopefulness,
and feelings of loneliness during the past week
(Costello & Devins, 1989; Radloff, 1977; Zich,
Atkison, & Greenfield, 1990).
Dot-Probe Paradigm. The computerized
Dot-Probe Paradigm was modified and pro-
grammed according to the specifications of
MacLeod et al. (1986). The program was run on
a Microsoft Windows XP PC and stimuli were
presented on a 17-in. monitor.
This task had 136 trials. There were eight
baseline-neutral practice words at the beginning
of the task, and 32 experimental word pairs
were then presented and repeated four times in
a computer-randomized order for each individ-
ual. The 32 experimental words were a part of
three trial conditions: (a) eight positive money
words (e.g., jackpot) with their matched con-
trols (e.g., “jasmine”), (b) eight negative money
words (e.g., debt) with their matched controls
(e.g., hunt), and (c) 16 neutral money words
(e.g., bank) with their matched controls (e.g.,
camp) (see Table 3). Each capitalized word was
ideographically matched for number of letters,
number of syllables, Kue`era–Francis written
frequency, concreteness rating and for its com-
mon part of speech (Wilson, 1987) and pre-
sented at random. Furthermore, control words
that would have a special meaning for students
(e.g., term) or that had a positive or negative
connotation (e.g., laughter,evicted), were ex-
cluded. The variation between the word groups
was explored and the means and standard devi-
ations were not found to differ beyond the
bounds of variability (see Table 4).
On each trial, the participant was first pre-
sented with a fixation point at the center of the
computer screen followed by a pair of words
presented for 500 milliseconds. The experimen-
tal words were presented at random. Once the
97MEASURING FINANCIAL ANXIETY
words simultaneously disappeared, a “dot
probe” appears and replaces one of the two
words. The difference in RTs when a dot probe
replaces the threat or neutral word either indi-
cates an avoidance or attentional capture (“vig-
ilance”) to the threat words (MacLeod et al.,
1986; Mogg, Bradley, Bono, & Painter, 1997).
The dot probe remained on the screen until the
participant responded or for a total of 4 seconds
if there was no response. Response latencies
were recorded to the nearest millisecond
through Visual Basic.
Procedure. After the participant read the
instructions for the DPP, the participants com-
pleted a practice trial, which was followed by the
probe detection task. The implicit DPP task was
conducted before the questionnaires were admin-
istered so that the reporting on one’s finances
would not prime participants’ DPP result. Partic-
ipants then completed the questionnaires. On av-
erage, the study was completed in 25 minutes.
Results: Data Reduction and Analysis
The data was analyzed using SPSS version 16.0
for MacOS X (SPSS Inc., 1989–2007).
Factor analysis on the FAS. An explor-
atory factor analysis was done to investigate
whether the different variables in the FAS
loaded onto the same underlying factor.
5
This
analysis suggested that two of the original items
included in the questionnaire were not associ-
ated with the same factor and slightly decreased
the reliability of the scale (see Table 2). This
was confirmed by observing the correlation ma-
trices. These two questions were therefore ex-
cluded from the FAS. The exclusion of these
items increased the Cronbach’s alpha negligibly
(␣⫽.850 to ␣⫽.855),
6
but more importantly,
reduced the components in the analysis (from
three to one). This also made theoretical sense
as the loading on question aseemed to be driven
by boredom while dseemed to be driven by
distrust, and both seemed weakly related to the
factor component measuring financial anxiety.
An exploratory factor analysis of the remain-
ing 10 items showed the scale was driven by a
single component, making it a unidimentional
scale.
7
A single financial anxiety score was
computed into a new variable.
Investigation the shared variation of the
FAS with depression and state anxiety.
The population had a mean score of 2.91 on
the FAS (SD ⫽0.65), of 15.24 on the CES-D
5
This test was “exploratory” because a factor analysis is
most reliable when there are at least 10–15 participants per
question or a sample of 300 or more, while the sample size
of this study (79 participants) did not satisfy either condi-
tion. However, a Kaiser-Meyer-Olkin measure of sampling
adequacy (KMO; a statistic that indicates the proportion of
variance in one’s variables that might be caused by under-
lying factors) was found to be between .8 and .9, indicating
that the patterns of correlations are reasonably compact and
that the factor analysis would likely yield reliable results
(Field, 2005). Also, a highly significant finding on the
Bartlett’s Test of Sphericity indicated that the variables in
this factor analysis are related and therefore suitable for
structure detection.
6
The excluded variables were the only two whose dele-
tion would increase Cronbach’s ␣.
7
Though only one component has been reported (with an
eigen value of 4.26), the eigen value for the second factor
was just above 1 (1.22). As noted above, a factor analysis
with just 79 cases will produce unstable eigen values, but
this is taken as evidence that there is not a strong case
against the unidimentionality of this solution.
Table 3
Words Used in DPP, by Condition and Their Respective Control Words
Positive words Negative words Neutral words
Positive
words Control
words Negative
words Control
words Neutral
words Control
words Neutral
words Control
words
Bonus Comet Credit Handle Bank Camp Financial Immediate
Jackpot Jasmine Debt Hunt Account Measure Salary Agency
Earnings Bathroom Unemployed Changeable Finance Suspect Cash Tape
Scholarship Measurement Loan Fill Money Woman Dollar Pocket
Bursary Caribou Bill Club Cost Turn Cheap Stiff
Grant Shift Bankrupt Converse Paid Hear Expensive Collected
Bingo Flour Mortgage Friction Economic Military Insurance Selection
Rich Fair Rent Jump Spending Describe Invoice Channel
98 SHAPIRO AND BURCHELL
(SD ⫽8.84), and of 37.08 on the STAI
(SD ⫽9.97).
The CES-D and STAI were significantly cor-
related, r⫽.721, p(two-tailed) ⬍.001. Finan-
cial anxiety (FAS) was significantly correlated
with low mood (CES-D), r⫽.461, and anxiety
(STAI), r⫽.388 (all ps (two-tailed)⬍.001). It
is important to note that first-order partial cor-
relations were conducted to address the overlap
in the variation of the scales and to determine
the size of the unique portion of variance (con-
trolling for the third scale).
8
The correlation
between the STAI and the FAS was not signif-
icantly related when the CES-D was controlled,
r⫽.091, ns. When controlling for anxiety
(STAI), there was a positive relationship be-
tween the FAS and CES-D, r⫽.283, p(two-
tailed) ⬍.05. Further, the partial correlation for
the CES-D and STAI (controlling for the FAS)
was r⫽.663, p(two-tailed) ⬍.001.
A linear regression model was conducted to
investigate how much the CES-D and STAI
predicted financial anxiety (see Table 5). The
assumption of no multicollinearity was tested
and was found to not have been violated (Tol-
erance ⫽.480). This model indicates that the
STAI and CES-D together can account
for 19.8% of the variance; however, only the
CES-D significantly predicts the FAS. This fur-
ther suggests that the correlation between the
FAS and CES-D is real, but the correlation
between the FAS and STAI is spurious.
Analysis of the DPP. A repeated-mea-
sures ANOVA was conducted to investigate
whether there was a significant difference by
word group, threat word position, and probe
position. Mauchly’s test indicated that the as-
sumption of sphericity had been violated for the
interaction effect between probe location and
threat word location,
2
(2) ⫽12.41, p⬍.01.
The degrees of freedom were corrected using
Greenhouse-Geisser estimates of sphericity
(ε⫽.86). There was a significant interaction
effect between probe location and threat loca-
tion, F(1, 72) ⫽4.075, p⬍.05, eta ⫽.23. This
indicates that RT toward the threat word dif-
fered depending on the subsequent probe place-
ment. Furthermore, there was no main effect of
threat location, probe location or group. There
was also no significant interaction effect be-
tween probe location ⫻group, threat location ⫻
group, or threat location ⫻probe location ⫻
group.
Avoidance or vigilance? For each word
group on the DPP (positive, negative, neutral, as
well as a combined group), the difference be-
tween the incongruent (threat word placed on
top while the probe was placed on bottom, or
visa versa) and congruent (threat and probe both
placed on the top or on the bottom) RTs were
calculated. This “direction” calculation com-
pares the RT of a participant on conditions
when the dot probe replaces the threat word
8
This was necessary because the STAI and CES-D have
a very large shared variation, R
2
⫽.52.
Table 4
Summary of Linguistic Property Variation (by Condition)
Number of letters Number of
syllables Written frequency Concreteness rating
NMean SD N Mean SDNMean SD N Mean SD
Positive words 8 6.50 2.27 8 2.00 .76 5 35.60 27.43 1 377.00 .
Positive control words 8 6.50 2.27 8 2.00 .76 6 30.00 27.42 1 413.00 .
Negative words 8 6.00 2.29 8 1.63 .74 8 39.25 46.73 5 435.00 57.34
Negative control words 8 6.00 2.39 8 1.63 .74 8 39.00 46.63 5 432.00 50.99
Neutral words 16 6.38 1.89 16 2.06 .93 15 98.60 83.42 11 458.91 104.23
Neutral control words 16 6.38 1.89 16 2.06 .93 16 87.81 75.00 12 457.58 98.49
Total words 64 6.31 2.04 64 1.94 .83 58 66.69 68.38 35 447.54 85.70
Table 5
Regression Model
Model B SE B
Constant 1.248 .807
CES-D .323 .125 .377
ⴱ
STAI .141 .177 .116
R
2
(adj) ⫽.198.
p⬍.001.
ⴱ
p⬍.05.
99MEASURING FINANCIAL ANXIETY
compared to when the dot probe replaces the
control word. A positive score indicates avoid-
ance whereas a negative score indicates vigi-
lance.
All scales showed an average of a positive
score, indicating that for most participants the
overall mechanism was avoidance of financial
words.
Relationship between the FAS and the
DPP. The negative financial word group in
the DPP calculations significantly correlated
with the FAS scale, ⫽.218, p (one-tailed)
⬍.05. This indicates that those who reported
having higher levels of financial anxiety (on the
FAS) display a greater RT delay on negative
threat words.
Discussion
The analysis of depression and general anxi-
ety indicated that though interrelated, the FAS
measure is a distinct construct from depression
(CES-D) and general anxiety (STAI). It is coun-
terintuitive that partial correlations indicated re-
latedness of the FAS with the CES-D, but not
the STAI. This suggests that low mood and
anxiety interacts with financial anxiety in dif-
ferent ways and while mood may relate directly,
anxiety seems to relate indirectly through mood.
Nevertheless, the finding that the FAS is distinct
from depression and general anxiety indicates
that the FAS is a useful tool in evaluating self-
reported financial anxiety in future studies.
The validity of the FAS is further enhanced
by the significant relationship that was found
between the FAS and the involuntary, nondirec-
tional DPP (on negative words). This finding
constructively extends the first study that inves-
tigated the relationship between self-reported
financial anxiety and the financially adapted
EST. It is interesting that similar to other anx-
ieties, a financial anxiety is associated with bi-
ased attention.
The relationship between the FAS and DPP
was only observed for negative words (i.e., not
positive or neutral financial word groups),
which is understandable as negative financial
words are most likely to provoke an anxious
avoidant response. However, though the EST
used mainly neutral words (which were presum-
ably less sensitive to picking up anxiety com-
pared to solely negative words), the EST still
resulted in a slightly stronger effect (r⫽.301)
than the negative words in the DPP (r⫽.218).
The slight difference in effect size between the
studies may be nonsignificant given the small
sample size in Study 1. However, this close
replication gives credibility to the measure-
ments used in these studies’ convergent ap-
proach.
This finding does not suggest that anxiety is
the only reason for financial avoidance. Indeed,
the correlation between a self-reported financial
anxiety and avoidance to financial stimuli only
had a medium effect size (r
2
⫽.09). This indi-
cates that while financial anxiety is important,
there are other factors at work (e.g., a lack of
interest in finances), which may need to be
considered in understanding financial avoid-
ance.
The DPP, a more sophisticated measure than
the EST (for reasons described above), also
provided insight into the mechanism of finan-
cial anxiety. On all the word groups (positive,
negative and neutral), avoidance (rather than
vigilance) was the most common strategy (even
on positive financial words, e.g., bonus). This
suggests that financial anxiety behaves like a
phobia, whereby individuals involuntarily at-
tempt to reduce their anxious mood state by
minimizing their encounter with fear-relevant
stimuli (Thorpe & Salkovkis, 1999). From a
functional perspective, similar to other anxieties
whereby avoidance is the predominant mecha-
nism, most individuals who are financially anx-
ious seemingly use avoidance as a defense
mechanism (Miu et al., 2008).
Future Research
This study developed novel instruments for
measuring financial anxiety. In so doing, this
study has answered but also raised questions.
Examining financial anxiety in a larger and di-
verse population would be useful in further in-
vestigating the demographics of financial anxi-
ety. As these studies were conducted in the U.K.
in 2008–9, further studies would determine how
generalizable these findings are across different
countries, as well as determine the effect that
the global economic recession may have had on
financial anxiety. Specifically, the inclusion of
age, income (or parental income), size and num-
ber of debts, and monthly budgets would be
beneficial to better ascertain how these variables
relate to financial anxiety.
100 SHAPIRO AND BURCHELL
Moreover, the results above are correlations
and do not imply causation. Accordingly, a lon-
gitudinal approach would be suitable to test
whether there is a causal link between financial
anxiety and debt by investigating whether fi-
nancial anxiety performance on the DPP could
predict subsequent debt (See Marissen et al.,
2006, who found performance on DPP to pre-
dict relapse in opiate-dependent individuals).
Though the correlations between the FAS
and the FEST and between the FAS and the
DPP indicate that the FAS contains a construct
validity, further psychometrics of the FAS must
be conducted on a larger sample size before the
FAS is widely accepted and utilized as a mea-
sure of self-reported financial anxiety.
Furthermore, similar to Amin, Constable, and
Canli (2004), in future neuroeconomic studies that
investigate financial anxiety it would be useful to
investigate brain activation (particularly of the
amygdala), galvanic skin response, as well as a
wider range of peripheral variables (e.g., cardio-
vascular activity, facial electromyography, sali-
vary cortisol) in conjuncture to the DPP and FEST
when possible (Dunn, Dalgleigh, & Lawrence,
2006). This would be informative in understand-
ing how financial anxiety is neurologically similar
to other phobias and how temporal and spatial
aspects of attentional mechanisms are registered.
Examining physiological responses of the general
population toward their personal finances would
usefully extend the work of Lo and Repin (2001),
who studied the physiological characteristics of
professional securities traders when they were en-
gaging in live trading. Furthermore, neurobiolog-
ical studies have strongly indicated that there is a
biological basis to trait anxiety, and some of the
genetic bases of trait anxiety have been identified
(Miu et al., 2008). It is likely that this line of
research will greatly advance our knowledge of
the relationship between state and trait anxiety, as
wellas the relationships between specific and non-
specific anxiety disorders.
As Miu et al. (2008) explain, trait anxiety has
been evaluated by examining attention (i.e.,
whether participant detect emotionally threatening
stimuli vs. neutral stimuli), interpretation (i.e.,
whether participants interpret stimuli as poten-
tially threatening), and memory (i.e., whether par-
ticipants favor the recall of information that is
related to the potential threat). This study has
examined financial anxiety in relation to attention;
however, future studies can extend upon this work
by evaluating how financial anxiety relates to in-
terpretation and memory of financial stimuli.
The evidence presented here suggests that fi-
nancial anxiety behaves like a phobia (as mea-
sured by an individual’s implicit reactions to fi-
nancial stimuli); this suggests that treatments that
have proven to be successful in the improvement
of other phobias, such as Cognitive Behavioral
Therapy, might be an effective way of helping
individuals with high levels of financial anxiety.
Further research would be necessary to evaluate
what may improve one’s financial anxiety.
Lastly, the integration of financial anxiety, atti-
tudes, behaviors, and decisions into a model of
student debt would be useful in explaining how
these factors are interrelated and which are espe-
cially problematic for student finances. The par-
ticular role of depression should be incorporated
into future work. This would require a meta-
analysis and model development (Robert & Jones,
2001).
Concluding Remarks
Financial anxiety can become devastating for
the individual as well as an obstacle to the goal
of governments in creating financially self-
sufficient citizens (Burchell, 2003). Personal fi-
nancial products continues to grow in number
and complexity; for instance, Internet loans
have recently become a big business, providing
fast loans with typical annual interest rates in
excess of 2,000%. It is in the interest of gov-
ernments to take financial anxieties seriously
and help citizens cope better with their personal
finances. For example, it is important to make
financial knowledge more accessible and finan-
cial products more consumer-friendly, which
may reduce individuals’ financial anxiety.
In both studies presented above, the self-
reported financial anxiety questionnaire corre-
lated significantly with the implicit measures.
These findings indicate that those who report
having financial anxiety also display reaction
latencies in the processing of financial informa-
tion. The development of these reported and
subliminal tools that measure financial anxiety
can be usefully employed in future research to
better understand financial anxiety. Accord-
ingly, financial behavior could be more compre-
hensively evaluated and policy could be better
determined by incorporating financial anxiety
101MEASURING FINANCIAL ANXIETY
into models of financial illiteracy, mismanage-
ment, and debt.
References
Ackert, L. F., Church, B. K., & Deaves, R. (2003). Emo-
tion and financial markets. Federal Reserve Bank of
Atlanta Economic Review, 88, 33–41. Retrieved from
http://www.frbatlanta.org/filelegacydocs/ackert_
q203.pdf
Amin, Z., Constable, R. T., & Canli., T. (2004).
Attentional bias for valenced stimuli as a function
of personality in the dot-probe task. Journal of
Research in Personality, 38, 15–23.
Amir, N., McNally, R. J., Riemann, B. C., Burns, J.,
Lorenz, M., & Mullen, J. T. (1996). Suppression of
the emotional Stroop effect by increased anxiety in
patients with social phobia. Behaviour Research
and Therapy, 34, 945–948.
Baker, P. M., & Hagedorn, R. B. (2008). Attitudes to
money in a random sample of adults: Factor anal-
ysis of the MAS and MBBS scales, and correla-
tions with demographic variables. Journal of So-
cio-Economics, 37, 1803–1814.
Baumler, G. (1985). Farbe-Wort-Interferenztest
(WIT) nach J. R. Stroop. Gottingen, Germany:
Verlag fur Psychologie Hogrefe.
Beal, D. J., & Delphachitra, S. B. (2003). Financial
literacy among Australian university students.
Economic Papers, 22, 65–78.
Burchell, B. J. (2002). Results of 300 telephone In-
terviews: Attitudes, knowledge and emotions con-
cerning personal finance. Unpublished report. Re-
trieved from http://people.pwf.cam.ac.uk/bb101/
SummaryofResultsof300TelephoneInterviews.pdf
Burchell, B. J. (2003). Identifying, describing and
understanding financial aversion: Financial
phobes. Report for EGG. Retrieved from http://
people.pwf.cam.ac.uk/bb101/FinancialAversion
ReportBurchell.pdf
Costello, C. G., & Devins, G. M. (1989). Screening
for depression among women attending their fam-
ily physicians. Canadian Journal of Behavioural
Science, 21, 434–451.
Damasio, A. R. (1994). Descartes error: Emotion,
reason, and the human brain. New York, NY:
Putnam.
Davies, E., & Lea, S. E. G. (1995). Student attitudes
to student debt. Journal of Economic Psychol-
ogy, 16, 663–679.
De Ruiter, C., & Brosschot, J. F. (1994). The emo-
tional Stroop interference effect in anxiety: Atten-
tional bias or cognitive avoidance? Behaviour Re-
search and Therapy, 32, 315–319.
Dunn, B. D., Dalgleish, T., & Lawrence, A. D.
(2006). The somatic marker hypothesis: A critical
evaluation. Neuroscience and Biobehavioral Re-
views, 30, 239–271.
Ehrman, R. N., Robbins, S. J., Bromwell, M. A.,
Lankford, M. E., Monterosso, J. R., & O’Brien,
C. P. (2002). Comparing attentional bias to smok-
ing cues in current smokers, former smokers, and
non-smokers using a dot-probe task. Drug and
Alcohol Dependence, 67, 185–191.
Faul, F., & Erdfelder, E. (1992). GPOWER: A priori,
post-hoc, and compromise power analyses for MS-
DOS [Computer program]. Bonn, FRG: Bonn Uni-
versity, Dept. of Psychology.
Field, A. (2005). Discovering statistics using SPSS.
London, UK: Sage.
Fox, E. (1993). Attentional bias in anxiety: Selective
or not? Behaviour Research and Therapy, 31, 487–
493.
Furnham, A. (1984). Many sides to the coin: The
psychology of money ssage. Personality and Indi-
vidual Differences, 5, 501–509.
Garman, E. T., Leech, I. E., & Grable, J. E. (1996).
The negative impact of employee poor personal
financial behaviors on employers. Financial Coun-
seling and Planning, 7, 157–168.
Hanoch, Y. (2002). Neither an angel nor an ant:
Emotion as an aid to bounded rationality. Journal
of Economic Psychology, 23, 1–25.
Hastings, J., & Tejeda-Ashton, L. (2008). Financial
literacy, information, and demand elasticity: Sur-
vey and experimental evidence from Mexico.
(NBER Working Paper 14538). Cambridge, MA:
National Bureau of Economic Research.
Henry, R. A., Weber, J. G., & Yarbrough, D. (2001,
June). Money management practices of college
students: Statistical data included. College Student
Journal, 7–14.
Jenkins, R., Bhugra, D., Bebbington, P., Brugha, T.,
Farrell, M., Coid, J., . .. Meltzer, H. (2008). Debt
income and mental disorder in the general popu-
lation. Psychological Medicine, 38, 1485–1493.
Joo, S., & Grable, J. E. (2000). Improving employee
productivity: The role of financial counseling and
education. Journal of Employment Counseling, 37,
2–15.
Jorgensen, B. L. (2007). Financial literacy of college
students: Parental and peer influences. (Unpub-
lished doctorial dissertation). Virginia Polytechnic
Institute and State University, Blacksburg, VA.
Jump$tart Coalition. (2004). 2004 Personal financial sur-
vey of high school seniors. Washington, DC: Jump$tart
Coalition for Personal Financial Literacy.
Kindt, M., Bierman, D., & Brosschot, J. F. (1996). Stroop
versus Stroop: Comparison of a card format and a
single-trial format of the Standard Color-Word Stroop
Task and the Emotional Stroop Task. Personality and
Individual Differences, 21, 653–661.
Lange, C., & Byrd, M. (1998). The relationship be-
tween perceptions of financial distress and feelings
102 SHAPIRO AND BURCHELL
of psychological well-being in New Zealand uni-
versity students. International Journal of Adoles-
cence and Youth, 7, 193–209.
Lim, V. K. G., & Sng, Q. S. (2006). Does parental job
insecurity matter? Money anxiety, money motives,
and work motivation. Journal of Applied Psychol-
ogy, 91, 1078–1087.
Lo, A. W., & Repin, D. V. (2001). The psychophys-
iology of real-time financial risk processing. Jour-
nal of Cognitive Neuroscience, 14, 323–339.
Lusardi, A., & Tufano, P. (2009). Debt literacy,
financial experiences, and overindebtedness.
(NBER Working Paper 14808). Cambridge, MA:
National Bureau of Economic Research.
MacLeod, C., Mathews, A., & Tata, P. (1986). At-
tentional bias in the emotional disorders. Journal
of Abnormal Psychology, 95, 15–20.
Marissen, M. A., Franken, I. H., Waters, A. J., Blan-
ken, P., van den Brink, W., & Hendriks, V. M.
(2006). Attentional bias predicts heroin relapse
following treatment. Addiction, 101, 1306–1312.
Medina, J. F., Saegert, J., & Gresham, A. (1996).
Comparison of Mexican-American and Anglo-
American attitudes toward money. The Journal of
Consumer Affairs, 30, 124–145.
Miu, A. C., Miclea, M., & Houser, D. (2008). Anx-
iety and decision-making: Toward a neuroeconom-
ics perspective. Advances in Health Economics
and Health Services Research, 20, 55–84.
Mogg, K., Bradley, B. P., Bono, J., & Painter, M.
(1997). Time course of attentional bias for threat
information in non-clinical anxiety. Behaviour Re-
search and Therapy, 35, 297–303.
Morgan, C. J., Rees, H., & Curran, H. V. (2008). Atten-
tional bias to incentive stimuli in frequent ketamine
users. Psychological Medicine, 38, 1331–1340.
National Council on Economic Education (NCEE).
2005. What American teens and adults know about
economics. Retrieved from http://207.124.141.218/
WhatAmericansKnowAboutEconomics042605–
3.pdf
Pflugshaupt, T., Mosimann, U. P., von Wartburg, R.,
Schmitt, W., Nyffeler, T., & Muri, R. M. (2005).
Hypervigilance-avoidance pattern in spider pho-
bia. Journal of Anxiety Disorders, 19, 105–116.
Radloff, L. S. (1977). The CES-D Scale: A self-
report depression scale for research in the general
population. Journal of Applied Psychological
Measures, 1, 385–401.
Ricciardi, V. (2008). The financial psychology of
worry and women. Working paper. Retrieved from
http://ssrn.com/abstract⫽1093351
Roberts, J. A., & Jones, E. (2001). Money attitudes,
credit card use, and compulsive buying among
American college students. Journal of Consumer
Affairs, 35, 213–240.
Roberts, R., Golding, J., Towell, T., & Weinreb, I.
(1999). The effects of economic circumstances on
British students’ mental and physical health. Jour-
nal of American College Health, 48, 103–109.
Rofey, D. L., Corcoran, K. J., & Tran, G. Q. (2004).
Bulimic symptoms and mood predict food relevant
Stroop interference in women with troubled eating
patterns. Eating Behaviors, 5, 35–45.
Rolls, E. T. (1999). The brain and emotion. Oxford,
UK: Oxford University Press.
Stetter, F., Ackermann, K., Bizer, A., Straube, E. R.,
& Mann, K. (1995). Effects of disease-related cues
in alcoholic inpatients: Results of a controlled “Al-
cohol Stroop” study. Alcoholism: Clinical and Ex-
perimental Research, 19, 593–599.
Stroop, J. R. (1935). Studies of interference in serial
verbal reactions. Journal of Experimental Psychol-
ogy, 18, 643–662.
Tata, P. R., Leibowitz, J. A., Prunty, M. J., Cameron,
M., & Pickering, A. D. (1996). Attentional bias in
obsessional compulsive disorder. Behaviour Re-
search and Therapy, 34, 53–60.
Thorpe, S. J., & Salkovkis, P. M. (1999). Animal
phobias. In D. C. L. Davey (Ed.), Phobias: A
handbook of theory, research and treatment.
Chichester, UK: Wiley.
Townshend, J. M., & Duka, T. (2001). Attentional
bias associated with alcohol cues: Differences be-
tween heavy and occasional social drinkers. Psy-
chopharmacology, 157, 67–74.
VanderZee, K. I., Sanderman, R., & Heyink, J.
(1996). A comparison of two multidimensional
measures of health status: The Nottingham Health
Profile and the RAND 36-Item Health Survey 1.0.
Quality of Life Research, 5, 165–174.
Wang, J., & Xiao, J. J. (2009). Buying behavior,
social support and credit card indebtedness of col-
lege students. International Journal of Consumer
Studies, 33, 2–10.
Warwick, J., & Mansfield, P. (2000). Credit card con-
sumers: College students’ knowledge and attitude.
Journal of Consumer Marketing, 17, 617–626.
Williams, J. M. G., Macleod, C., & Mathews, A.
(1996). The Emotional Stroop Task and psycho-
pathology. Psychological Bulletin, 120, 3–24.
Wilson, M. (1987). MRC Psycholinguistic Database:
Machine Usable Dictionary. Version 2.00. Chilton,
UK: Informatics Division Science and Engineering
Research Council Rutherford Appleton Laboratory.
Yamauchi, K., & Templer, D. (1982). The develop-
ment of a money attitudes scale. Journal of Per-
sonality Assessment, 46, 522–528.
Zich, J. M., Atkison, C. C., & Greenfield, T. K.
(1990). Screening for depression in primary care
clinics: The CES-D and the BDI. International
Journal of Psychiatry of Medicine, 20, 259–277.
Received October 21, 2011
Revision received January 25, 2012
Accepted January 31, 2012 䡲
103MEASURING FINANCIAL ANXIETY
Copyright of Journal of Neuroscience, Psychology, & Economics is the property of Association for
NeuroPsychoEconomics and its content may not be copied or emailed to multiple sites or posted to a listserv
without the copyright holder's express written permission. However, users may print, download, or email
articles for individual use.
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Neuroscience Psychology and Economics
This content is subject to copyright. Terms and conditions apply.
